We use HPLC extensively to determine analyte concentrations in (often) complex matrices. I am interested in people's view of the practice of using injector volume variation to create a calibration curve from a single standard preparation (as opposed to making up the standards over a range of concentrations and injecting each at a constant volume).
My own view is that it is far preferable to create a "real" calibration curve over a range of concentrations. Opinion, anyone?

I agree with you Lorin, when you only have a single standard preperation and inject different volumes, you won't get any control of you standard preperation.
When you have a range of concentrations and inject the same volume, you'll be able to control your standard.

But how about a double standard preperation, where you'll inject the first (lets say SA) five times, and use the second (SB) to control your SA ?
You have to have a good autosampler precision to do this

Don't know what Holger is on about! surely we have had enough of double standards????
With variable injection, precision around 1% is possible, with fixed loop injection, 0.1% is possible, with automated variable loop about 0.5% is possible. If you only make up one standard, you are relying on that preparation to be accurate, if you make up several, then slight variation maybe acceptable. Do you run analytical quality control checks, everyone should!
I suppose it depends on the work you are doing, and the accuracy/precision you are aimimg for,.... and how good are you at injecting?

In an ideal world, the best results should be obtained by making serial dilutions from a stock solution in mobil phase and injecting a constant volume. As we all know this is a pain to do and sometimes not even possible. If you want (or need) to use variable volume injection calibrations you would need to do is show that the results you obtain from individal standards are not highly different than your variable volume injections. Assuming that the differences between the two methods are acceptable, there would be little reason not to use variable volumes. It should be pointed out this experimant would need to be done for each different assay and should not be considered as a global solution.

I guess like everything we do "it depends". It depends on what the expected concentration range is, the design of your equipment, how your equipment is calibrated/validated, how your methods are developed/validated, etc.

With modern equipment(HP1100s)we validate both response linearity using different concentrations and injector linearity using different injection volumes.

I have no qualms with developing more efficient methods that calibrate using mass on column from different volume injections. Gather some data and look at the precision, linearity, and specific response.

Cheers and Good Luck,

Tom

I ran a comparison between conventional calibration when same volume was injected from the solutions of different concentrations vs. volume injection range and did not find any difference. Was it a validation? Probably, for this project.

An interesting extension of this query is to ask if anyone has looked at the precision on letting the autosampler make dilutions for those new autosamplers that have that capability? I have one but have never used that capability - its one of those projects that would be fun to do "when I have a little extra time"